Re: [Qemu-devel] [PATCH v3 2/2] qdev: Check for the availability of a hotplug controller before adding a device

[Markus Armbruster]

Thomas Huth <address@hidden> writes:

> The qdev_unplug() function contains a g_assert(hotplug_ctrl) statement,
> so QEMU crashes when the user tries to device_add + device_del a device
> that does not have a corresponding hotplug controller. This could be
> provoked for a couple of devices in the past (see commit 4c93950659487c7ad
> or 84ebd3e8c7d4fe955 for example), and can currently for example also be
> triggered like this:
>
> $ s390x-softmmu/qemu-system-s390x -M none -nographic 
> QEMU 2.10.50 monitor - type 'help' for more information
> (qemu) device_add qemu-s390x-cpu,id=x
> (qemu) device_del x
> **
> ERROR:qemu/qdev-monitor.c:872:qdev_unplug: assertion failed: (hotplug_ctrl)
> Aborted (core dumped)
>
> So devices clearly need a hotplug controller when they should be usable
> with device_add.
> The code in qdev_device_add() already checks whether the bus has a proper
> hotplug controller,

Where?  Hmm, I guess it's this one:

    if (qdev_hotplug && bus && !qbus_is_hotpluggable(bus)) {
        error_setg(errp, QERR_BUS_NO_HOTPLUG, bus->name);
        return NULL;
    }

>                     but for devices that do not have a corresponding bus,
> there is no appropriate check available yet. In that case we should check
> whether the machine itself provides a suitable hotplug controller and
> refuse to plug the device if none is available.
>
> Reviewed-by: Igor Mammedov <address@hidden>
> Signed-off-by: Thomas Huth <address@hidden>
> ---
>  hw/core/qdev.c         | 28 ++++++++++++++++++++--------
>  include/hw/qdev-core.h |  1 +
>  qdev-monitor.c         |  5 +++++
>  3 files changed, 26 insertions(+), 8 deletions(-)
>
> diff --git a/hw/core/qdev.c b/hw/core/qdev.c
> index 1111295..f739753 100644
> --- a/hw/core/qdev.c
> +++ b/hw/core/qdev.c
> @@ -253,19 +253,31 @@ void qdev_set_legacy_instance_id(DeviceState *dev, int 
> alias_id,
>      dev->alias_required_for_version = required_for_version;
>  }
>  
> +HotplugHandler *qdev_get_machine_hotplug_handler(DeviceState *dev)
> +{
> +    MachineState *machine;
> +    MachineClass *mc;
> +    Object *m_obj = qdev_get_machine();
> +
> +    if (object_dynamic_cast(m_obj, TYPE_MACHINE)) {
> +        machine = MACHINE(m_obj);
> +        mc = MACHINE_GET_CLASS(machine);
> +        if (mc->get_hotplug_handler) {
> +            return mc->get_hotplug_handler(machine, dev);
> +        }
> +    }
> +
> +    return NULL;
> +}
> +
>  HotplugHandler *qdev_get_hotplug_handler(DeviceState *dev)
>  {
> -    HotplugHandler *hotplug_ctrl = NULL;
> +    HotplugHandler *hotplug_ctrl;
>  
>      if (dev->parent_bus && dev->parent_bus->hotplug_handler) {
>          hotplug_ctrl = dev->parent_bus->hotplug_handler;
> -    } else if (object_dynamic_cast(qdev_get_machine(), TYPE_MACHINE)) {
> -        MachineState *machine = MACHINE(qdev_get_machine());
> -        MachineClass *mc = MACHINE_GET_CLASS(machine);
> -
> -        if (mc->get_hotplug_handler) {
> -            hotplug_ctrl = mc->get_hotplug_handler(machine, dev);
> -        }
> +    } else {
> +        hotplug_ctrl = qdev_get_machine_hotplug_handler(dev);
>      }
>      return hotplug_ctrl;
>  }

qdev_get_machine_hotplug_handler() factored out of
qdev_get_hotplug_handler().  Okay.  Announcing it in the commit message
could've saved me a few review brainwaves.

> diff --git a/include/hw/qdev-core.h b/include/hw/qdev-core.h
> index 0a71bf8..51473ee 100644
> --- a/include/hw/qdev-core.h
> +++ b/include/hw/qdev-core.h
> @@ -286,6 +286,7 @@ DeviceState *qdev_try_create(BusState *bus, const char 
> *name);
>  void qdev_init_nofail(DeviceState *dev);
>  void qdev_set_legacy_instance_id(DeviceState *dev, int alias_id,
>                                   int required_for_version);
> +HotplugHandler *qdev_get_machine_hotplug_handler(DeviceState *dev);
>  HotplugHandler *qdev_get_hotplug_handler(DeviceState *dev);
>  void qdev_unplug(DeviceState *dev, Error **errp);
>  void qdev_simple_device_unplug_cb(HotplugHandler *hotplug_dev,
> diff --git a/qdev-monitor.c b/qdev-monitor.c
> index 9188d20..38c0fc2 100644
> --- a/qdev-monitor.c
> +++ b/qdev-monitor.c
> @@ -614,6 +614,11 @@ DeviceState *qdev_device_add(QemuOpts *opts, Error 
> **errp)
>  
       if (qdev_hotplug && bus && !qbus_is_hotpluggable(bus)) {
           error_setg(errp, QERR_BUS_NO_HOTPLUG, bus->name);
           return NULL;
       }

       if (!migration_is_idle()) {
           error_setg(errp, "device_add not allowed while migrating");
           return NULL;
       }

       /* create device */
       dev = DEVICE(object_new(driver));

>      if (bus) {
>          qdev_set_parent_bus(dev, bus);
> +    } else if (qdev_hotplug && !qdev_get_machine_hotplug_handler(dev)) {
> +        /* No bus, no machine hotplug handler --> device is not hotpluggable 
> */

Long line.

> +        error_setg(&err, "Device '%s' can not be hotplugged on this machine",
> +                   driver);
> +        goto err_del_dev;
>      }
>  
>      qdev_set_id(dev, qemu_opts_id(opts));

Hmm.  We need to check "can hotplug" in two separate ways, with bus and
without bus.  Can we keep the two ways on one place?  Something like

       if (qdev_hotplug) {
           if (bus && !qbus_is_hotpluggable(bus)) {
               error_setg(errp, QERR_BUS_NO_HOTPLUG, bus->name);
               return NULL;
           }
           if (!bus && !qdev_get_machine_hotplug_handler(dev)) {
               error_setg(&err,
                          "Machine doesn't support hot-plugging device '%s'"
                          driver);
               return NULL;
           }
      }